Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Making the Ciphertext String Itself Truly Random – The Conjecture.
7 views
Skip to first unread message
adacrypt
Feb 12, 2012, 7:30:30 AM2/12/12
to
I have made the point recently that the ciphertext string is never
random and indeed the clever cryptanalyst who knows this will try and
determine to what extent that is instantaneously true so that he may
mount a statistical mapping attack using that as the basis (an
advanced Kasiski/Babbage attack).
My ploy of studiously making the ciphertext string truly random or at
least very close to that is the appropriate foil to this last ditch
attack by the cryptanalyst.
I claim that this is an innovative new approach in cryptography.
Bearing in mind that all current number-theoretic cryptography
ridiculously puts the plaintext and the cipher text as neighbours on
the same straight line and the only thing saving the ciphertext from
becoming totally transparent is the disparity (there for the finding
also by the same cryptanalyst) of the relative addresses on the number
line of these two.
No. 1 attack is the inversion attack or ciphertext only attack, let
the next attack be indexed No 2 attack and is based on numerical
linear analysis and linear differential analysis using ordinary
mathematics.
The final attack No 3 seeks to circumvent both of these i.e. No’s 1 ,
2 in one fell swoop by attempting to map the ciphertext directly to
the plaintext it represents on a statistical basis. My answer to that
is to make the ciphertext truly random i.e. having no repeats so that
each element of ciphertext has equal probability of being the right
one and the attack then collapses into total uncertainty.
Having put so much thought into the cipher algorithm fearing
ciphertext only attack as the most likely one and to which the cipher
is most vulnerable, the cryptographer may forget or underestimate or
even be totally unable, to foil this third attack.
I want to address this last attack here and describe how it is
counteracted in my cryptography.
First of all the highly transparent number line situation where the
ciphertext is a sitting duck waiting to be shot at by a cryptanalyst
is foiled by the concept of ‘displacement’ where by the numbers that
represent the ciphertext are scattered wildly throughout three-
dimensional space instead of being addresses in the same street so to
speak.
Attack No 3 is foiled by making the ciphertext scientifically random.
This has to be structured from the very beginning in the cipher design
by a multiplicity of component keys only one of which must be truly
random i.e. the finished ciphertext string.
It is interesting to note that when a number of truly random keysets
i.e. having zero repeats, are factors in an encryption process they
lose their randomness when they operate on each other during the
process i.e. they can surface with such copious repeats that by the
definition they are not now random any more. This securing randomness
ploy needs to be prodigiously crafted into the ciphertext by the
design algorithm.
At the present time I have found by means of casual operational
research that the number of keysets should be three or four at least
and the numeric ciphertext needs to be as large as possible with seven
digits per number being a minimum. Clearly, such a rule-of-thumb
needs to be ratified more scientifically (deserves proper research)
but the interesting thing is that there is a new area for research
here and also, this obviates the need for generating separate random
keys beforehand something that has preoccupied cryptographers for over
a century. The nub of the matter seems to be the encryption algorithm
generates its own randomness internally without needing difficult-to-
find dedicated random keys to be provided from the outside as a
precondition.
There is a price to be paid for large-number ciphertext of course but
so be it if it means secure communications that are independent of
computer power for all time.
Recapping on the displacement cipher being called “Skew Line
Encryptions”
1) This form of attack i.e. ciphertext only attack is prevented by a
mathematical one-way function derived from a change-of origin being
given to the position vector that comprises the ciphertext.
2) This form of attack is foiled by the ciphertext string being made a
discontinuous string of wildly disparate elements.
3) This form of attack is foiled by the ciphertext string being made
truly random right through the creation stage.
This post is made with the best intentions, I am merely reporting back
to you things I have found while working on my own cryptography that I
think might be of interest to others. I am not being patronising or
pontificating to anybody.
- adacrypt
random and indeed the clever cryptanalyst who knows this will try and
determine to what extent that is instantaneously true so that he may
mount a statistical mapping attack using that as the basis (an
advanced Kasiski/Babbage attack).
My ploy of studiously making the ciphertext string truly random or at
least very close to that is the appropriate foil to this last ditch
attack by the cryptanalyst.
I claim that this is an innovative new approach in cryptography.
Bearing in mind that all current number-theoretic cryptography
ridiculously puts the plaintext and the cipher text as neighbours on
the same straight line and the only thing saving the ciphertext from
becoming totally transparent is the disparity (there for the finding
also by the same cryptanalyst) of the relative addresses on the number
line of these two.
No. 1 attack is the inversion attack or ciphertext only attack, let
the next attack be indexed No 2 attack and is based on numerical
linear analysis and linear differential analysis using ordinary
mathematics.
The final attack No 3 seeks to circumvent both of these i.e. No’s 1 ,
2 in one fell swoop by attempting to map the ciphertext directly to
the plaintext it represents on a statistical basis. My answer to that
is to make the ciphertext truly random i.e. having no repeats so that
each element of ciphertext has equal probability of being the right
one and the attack then collapses into total uncertainty.
Having put so much thought into the cipher algorithm fearing
ciphertext only attack as the most likely one and to which the cipher
is most vulnerable, the cryptographer may forget or underestimate or
even be totally unable, to foil this third attack.
I want to address this last attack here and describe how it is
counteracted in my cryptography.
First of all the highly transparent number line situation where the
ciphertext is a sitting duck waiting to be shot at by a cryptanalyst
is foiled by the concept of ‘displacement’ where by the numbers that
represent the ciphertext are scattered wildly throughout three-
dimensional space instead of being addresses in the same street so to
speak.
Attack No 3 is foiled by making the ciphertext scientifically random.
This has to be structured from the very beginning in the cipher design
by a multiplicity of component keys only one of which must be truly
random i.e. the finished ciphertext string.
It is interesting to note that when a number of truly random keysets
i.e. having zero repeats, are factors in an encryption process they
lose their randomness when they operate on each other during the
process i.e. they can surface with such copious repeats that by the
definition they are not now random any more. This securing randomness
ploy needs to be prodigiously crafted into the ciphertext by the
design algorithm.
At the present time I have found by means of casual operational
research that the number of keysets should be three or four at least
and the numeric ciphertext needs to be as large as possible with seven
digits per number being a minimum. Clearly, such a rule-of-thumb
needs to be ratified more scientifically (deserves proper research)
but the interesting thing is that there is a new area for research
here and also, this obviates the need for generating separate random
keys beforehand something that has preoccupied cryptographers for over
a century. The nub of the matter seems to be the encryption algorithm
generates its own randomness internally without needing difficult-to-
find dedicated random keys to be provided from the outside as a
precondition.
There is a price to be paid for large-number ciphertext of course but
so be it if it means secure communications that are independent of
computer power for all time.
Recapping on the displacement cipher being called “Skew Line
Encryptions”
1) This form of attack i.e. ciphertext only attack is prevented by a
mathematical one-way function derived from a change-of origin being
given to the position vector that comprises the ciphertext.
2) This form of attack is foiled by the ciphertext string being made a
discontinuous string of wildly disparate elements.
3) This form of attack is foiled by the ciphertext string being made
truly random right through the creation stage.
This post is made with the best intentions, I am merely reporting back
to you things I have found while working on my own cryptography that I
think might be of interest to others. I am not being patronising or
pontificating to anybody.
- adacrypt
adacrypt
Feb 12, 2012, 7:50:52 AM2/12/12
to
numbers are the ciphertext - even MM noticed this and is to be
commended for it - pity he has'nt a bit more brains - he's not a bad
lad for all that ! - adacrypt
biject
Feb 12, 2012, 12:58:55 PM2/12/12
to
On Feb 12, 5:30 am, adacrypt <austin.oby...@hotmail.com> wrote:
> The final attack No 3 seeks to circumvent both of these i.e. No’s 1 ,
> 2 in one fell swoop by attempting to map the ciphertext directly to
> the plaintext it represents on a statistical basis. My answer to that
> is to make the ciphertext truly random i.e. having no repeats so that
> each element of ciphertext has equal probability of being the right
> one and the attack then collapses into total uncertainty.
>
This seems nonsense. You give me a file that you think is random
> The final attack No 3 seeks to circumvent both of these i.e. No’s 1 ,
> 2 in one fell swoop by attempting to map the ciphertext directly to
> the plaintext it represents on a statistical basis. My answer to that
> is to make the ciphertext truly random i.e. having no repeats so that
> each element of ciphertext has equal probability of being the right
> one and the attack then collapses into total uncertainty.
>
that
has NO repeats at all and make the file at least 100,000 bytes
and I will supply a program to the Globe guy and it will with out
knowing the file in advance make it at least 100 bytes shorter
sense a file known to have no repeats is not RANDOM it can be
compressed to a smaller size by a general bijective file compressor.
And yes I will supply the decompressor to.
I hope you will learn anything from this but to think such a
file is random is to not understand what random is in my estimation.
But I think this was a weakness in enigma to during WWII.
> Having put so much thought into the cipher algorithm fearing
> ciphertext only attack as the most likely one and to which the cipher
> is most vulnerable, the cryptographer may forget or underestimate or
> even be totally unable, to foil this third attack.
>
ciphertext only attack is numro uno in that if a cipher can't stand
up to that its no good. But thats where it ends. I am not sure yours
can stand up to a ciphertext only attack especially since you have
no idea what a random file is.
At one time I though the ciphertext only attack was almost
everything
I was a slow learner. But Paul Onions who use to post on this site
showed weakness to a plain text attack in scott16 at first I ignored
it.
But as time went on I LEARNED I WAS WRONG so I made scott19u but it
was
to late for my credibility in crypto. And in the same sense it may
be to late for you since you current credibility might be low. Once
its low few people will not take anything you say about crypt
seriously.
Worse you may actually create something good someday if your capable
of learning here. But at that point the crypto gods my dismiss it with
out looking at it or declare it worthless and they may even discusss
attacks they think would make it fail even if such attacks don't.
This may sound impossible but it can happen. Look learn and try to be
friendly to those that help you. Its possible you could contribute
something but you have to LEARN FIRST.
You could also put you talents to writting compression code. The
nice thing about compression code it that its easier to test than
encryption so its obvious immediately if it works as advertised.
One point I would hope you do especially if you continue to write
cypto code and if you use several passes through the data try
using BWTS in one of the internal passes the big guys don't like it
and it greatly increases the unicity distance of the ciphertext if
done rightly and would not increase the length of output.
....
David A. Scott
--
My Crypto code
http://bijective.dogma.net/crypto/scott19u.zip
http://www.jim.com/jamesd/Kong/scott19u.zip old version
My Compression code http://bijective.dogma.net/
**TO EMAIL ME drop the roman "five" **
Disclaimer:I am in no way responsible for any of the statements
made in the above text. For all I know I might be drugged.
As a famous person once said "any cryptograhic
system is only as strong as its weakest link"
biject
Feb 12, 2012, 1:15:18 PM2/12/12
to
On Feb 12, 10:58 am, biject <biject.b...@gmail.com> wrote:
> This seems nonsense. You give me a file that you think is random
> that
> has NO repeats at all and make the file at least 100,000 bytes
> and I will supply a program to the Globe guy and it will with out
> knowing the file in advance make it at least 100 bytes shorter
I goofed again. I realized I could save at least 100 bits not bytes
> This seems nonsense. You give me a file that you think is random
> that
> has NO repeats at all and make the file at least 100,000 bytes
> and I will supply a program to the Globe guy and it will with out
> knowing the file in advance make it at least 100 bytes shorter
and since we right in byte I will write code to save at least 13
bytes.
the 100 bits was conservative so 13 bytes saves should be ok SORRY.
> sense a file known to have no repeats is not RANDOM it can be
> compressed to a smaller size by a general bijective file compressor.
> And yes I will supply the decompressor to.
> I hope you will learn anything from this but to think such a
> file is random is to not understand what random is in my estimation.
> But I think this was a weakness in enigma to during WWII.
>
Mark Murray
Feb 12, 2012, 1:42:35 PM2/12/12
to
On 12/02/2012 12:50, adacrypt wrote:
> In essence the cipher is a perfect random number generator wherein the
> numbers are the ciphertext - even MM noticed this and is to be
> commended for it - pity he has'nt a bit more brains - he's not a bad
> lad for all that ! - adacrypt
You are misquoting me again - badly. I never said that it was a perfect
> In essence the cipher is a perfect random number generator wherein the
> numbers are the ciphertext - even MM noticed this and is to be
> commended for it - pity he has'nt a bit more brains - he's not a bad
> lad for all that ! - adacrypt
RNG, in fact I was pointing out that you simply had a polyalphabetic
cipher, and I showed how to recofer the dictionaries.
Encipher a big file, then plot the vector components against each other
(X vs y), (y vs z) and (z vs x). Notice that there is strong structure
there!
Then note that all the A's encipher to ciphertext_A, all B's to
ciphertext_B where there is no intersection between these sets
at all. Thus, a dictionary is formed, and your cipher is busted.
M
--
Mark "No Nickname" Murray
Notable nebbish, extreme generalist.
Mark Murray
Feb 12, 2012, 1:48:47 PM2/12/12
to
On 12/02/2012 17:58, biject wrote:
> This seems nonsense. You give me a file that you think is random
> that
> has NO repeats at all and make the file at least 100,000 bytes
> and I will supply a program to the Globe guy and it will with out
> knowing the file in advance make it at least 100 bytes shorter
> sense a file known to have no repeats is not RANDOM it can be
> compressed to a smaller size by a general bijective file compressor.
> And yes I will supply the decompressor to.
> I hope you will learn anything from this but to think such a
> file is random is to not understand what random is in my estimation.
> But I think this was a weakness in enigma to during WWII.
Adacrypt's definition of random appears to be "has no repeats".
> This seems nonsense. You give me a file that you think is random
> that
> has NO repeats at all and make the file at least 100,000 bytes
> and I will supply a program to the Globe guy and it will with out
> knowing the file in advance make it at least 100 bytes shorter
> sense a file known to have no repeats is not RANDOM it can be
> compressed to a smaller size by a general bijective file compressor.
> And yes I will supply the decompressor to.
> I hope you will learn anything from this but to think such a
> file is random is to not understand what random is in my estimation.
> But I think this was a weakness in enigma to during WWII.
His mind is about as closed as its ever been when trying to shift
this misconception from his understanding.
Information theory, Shannon Entropy and the meaning of compressibility
are so far out of his realm of understanding that its not worth
bothering with. Best to use concepts no harder than high school
level.
Mark Murray
Feb 12, 2012, 2:40:55 PM2/12/12
to
On 12/02/2012 12:30, adacrypt wrote:
> 1) This form of attack i.e. ciphertext only attack is prevented by a
> mathematical one-way function derived from a change-of origin being
> given to the position vector that comprises the ciphertext.
You don't have a one-way function!
> 1) This form of attack i.e. ciphertext only attack is prevented by a
> mathematical one-way function derived from a change-of origin being
> given to the position vector that comprises the ciphertext.
> 2) This form of attack is foiled by the ciphertext string being made a
> discontinuous string of wildly disparate elements.
the same sets, and every element in the output set was equally likely
for any input element, then I could not have done this. What you are
dogmatically repeating is a massive weakness, not a strength.
> 3) This form of attack is foiled by the ciphertext string being made
> truly random right through the creation stage.
Oh, and learn what "random" means. It does not mean "no repeats".
> This post is made with the best intentions, I am merely reporting back
> to you things I have found while working on my own cryptography that I
> think might be of interest to others. I am not being patronising or
> pontificating to anybody.
You are now repeating a dogma. These are not findings, they are
desparate opinions. No criticism of them is accepted, considered
or even tolerated.
You have lost the plot.
0 new messages